Skip to content

ElasticSearch logging in RemoteRolloutProcessor (part 1)#240

Merged
dphuang2 merged 7 commits intomainfrom
rollout-logs-integration
Oct 1, 2025
Merged

ElasticSearch logging in RemoteRolloutProcessor (part 1)#240
dphuang2 merged 7 commits intomainfrom
rollout-logs-integration

Conversation

@dphuang2
Copy link
Copy Markdown
Collaborator

@dphuang2 dphuang2 commented Oct 1, 2025

  • initial setup handling done
  • python logging handler tested
  • TODO: propagate rollout metadata
  • TODO: example of how to propagate logs/status

Dylan Huang added 7 commits September 30, 2025 16:10
- Introduced setup method in RemoteRolloutProcessor to initialize Elasticsearch if not disabled.
- Added ElasticSearchConfig model for managing Elasticsearch configuration.
- Implemented logic to parse environment variables from a .env file and start Elasticsearch if necessary.
- Updated evaluation_test to call rollout_processor.setup() for proper initialization.
- Modified RolloutProcessor to include a setup method for potential overrides in subclasses.
- Removed the previous logic for parsing environment variables and starting Elasticsearch directly.
- Introduced a dedicated ElasticsearchSetup module to handle Elasticsearch initialization.
- Updated the _setup_elastic_search method to utilize the new setup module for improved clarity and maintainability.
- Renamed and updated methods for clarity: _setup_local_elasticsearch to _setup_existing_docker_elasticsearch and _setup_remote_elasticsearch to _setup_initialized_docker_elasticsearch.
- Improved comments to better describe the purpose of each setup method.
- Enhanced the logic for initializing Elasticsearch with Docker, ensuring clearer handling of existing and new setups.
- Introduced a new attribute 'pid' to langfuse_row to facilitate detection of stopped evaluations.
- Updated comments to clarify the purpose of the new attribute in relation to logging status updates.
- Introduced ElasticsearchDirectHttpHandler for asynchronous logging to Elasticsearch.
- Added ElasticsearchIndexManager for managing index creation and mapping configuration.
- Updated ElasticsearchSetup to create logging indices with proper mappings.
- Enhanced ElasticSearchConfig model to include index_name attribute.
- Implemented tests for ElasticsearchDirectHttpHandler to verify log transmission and sorting.
@dphuang2 dphuang2 changed the title Integrate ElasticSearch setup in RemoteRolloutProcessor ElasticSearch logging in RemoteRolloutProcessor Oct 1, 2025
@dphuang2 dphuang2 changed the title ElasticSearch logging in RemoteRolloutProcessor ElasticSearch logging in RemoteRolloutProcessor (part 1) Oct 1, 2025
@dphuang2 dphuang2 merged commit 3dc172c into main Oct 1, 2025
7 checks passed
@dphuang2 dphuang2 deleted the rollout-logs-integration branch October 1, 2025 18:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant